2 % (c) The GRASP/AQUA Project, Glasgow University, 1997
4 \section{String buffers}
6 Buffers for scanning string input stored in external arrays.
10 {-# OPTIONS -fno-prune-tydecls #-}
11 -- Don't really understand this!
12 -- ERROR: Can't see the data constructor(s) for _ccall_/_casm_ argument;
13 -- type: ForeignObj(try compiling with -fno-prune-tydecls ..)
21 hGetStringBuffer, -- :: FilePath -> IO StringBuffer
22 freeStringBuffer, -- :: StringBuffer -> IO ()
25 currentChar, -- :: StringBuffer -> Char
26 currentChar#, -- :: StringBuffer -> Char#
27 indexSBuffer, -- :: StringBuffer -> Int -> Char
28 indexSBuffer#, -- :: StringBuffer -> Int# -> Char#
29 -- relative lookup, i.e, currentChar = lookAhead 0
30 lookAhead, -- :: StringBuffer -> Int -> Char
31 lookAhead#, -- :: StringBuffer -> Int# -> Char#
33 -- moving the end point of the current lexeme.
34 setCurrentPos#, -- :: StringBuffer -> Int# -> StringBuffer
35 incLexeme, -- :: StringBuffer -> StringBuffer
36 decLexeme, -- :: StringBuffer -> StringBuffer
38 -- move the start and end lexeme pointer on by x units.
39 stepOn, -- :: StringBuffer -> StringBuffer
40 stepOnBy#, -- :: StringBuffer -> Int# -> StringBuffer
41 stepOnTo#, -- :: StringBuffer -> Int# -> StringBuffer
42 stepOnUntil, -- :: (Char -> Bool) -> StringBuffer -> StringBuffer
43 stepOverLexeme, -- :: StringBuffer -> StringBuffer
44 scanNumLit, -- :: Int -> StringBuffer -> (Int, StringBuffer)
45 expandWhile, -- :: (Char -> Bool) -> StringBuffer -> StringBuffer
46 expandUntilMatch, -- :: StrinBuffer -> String -> StringBuffer
47 -- at or beyond end of buffer?
48 bufferExhausted, -- :: StringBuffer -> Bool
49 emptyLexeme, -- :: StringBuffer -> Bool
52 prefixMatch, -- :: StringBuffer -> String -> Bool
53 untilEndOfString#, -- :: StringBuffer -> Int#
54 untilEndOfChar#, -- :: StringBuffer -> Int#
55 untilChar#, -- :: StringBuffer -> Char# -> Int#
58 lexemeToString, -- :: StringBuffer -> String
59 lexemeToByteArray, -- :: StringBuffer -> _ByteArray Int
60 lexemeToFastString, -- :: StringBuffer -> FastString
61 lexemeToBuffer, -- :: StringBuffer -> StringBuffer
67 #include "HsVersions.h"
70 import Addr ( Addr(..) )
73 import IO ( openFile, hFileSize, hClose, IOMode(..) )
75 #if __GLASGOW_HASKELL__ < 301
76 import IOBase ( IOError(..), IOErrorType(..) )
77 import IOHandle ( readHandle, writeHandle, filePtr )
78 import PackBase ( unpackCStringBA )
80 import PrelIOBase ( IOError(..), IOErrorType(..) )
81 import PrelHandle ( readHandle, writeHandle, filePtr )
82 import PrelPack ( unpackCStringBA )
94 -- ForeignObj# -- the data
101 instance Text StringBuffer where
102 showsPrec _ s = showString ""
106 hGetStringBuffer :: FilePath -> IO StringBuffer
107 hGetStringBuffer fname =
108 openFile fname ReadMode >>= \ hndl ->
109 hFileSize hndl >>= \ len@(J# _ _ d#) ->
110 let len_i = fromInteger len in
111 -- Allocate an array for system call to store its bytes into.
112 -- ToDo: make it robust
113 -- trace (show ((len_i::Int)+1)) $
114 _casm_ `` %r=(char *)malloc(sizeof(char)*(int)%0); '' (len_i::Int) >>= \ arr@(A# a#) ->
115 if addr2Int# a# ==# 0# then
116 fail (userError ("hGetStringBuffer: Could not allocate "++show len_i ++ " bytes"))
119 -- _casm_ `` %r=NULL; '' >>= \ free_p ->
120 -- makeForeignObj arr free_p >>= \ fo@(_ForeignObj fo#) ->
121 readHandle hndl >>= \ hndl_ ->
122 writeHandle hndl hndl_ >>
123 let ptr = filePtr hndl_ in
124 _ccall_ fread arr (1::Int) len_i ptr >>= \ (I# read#) ->
126 if read# ==# 0# then -- EOF or some other error
127 fail (userError ("hGetStringBuffer: failed to slurp in interface file "++fname))
129 -- Add a sentinel NUL
130 _casm_ `` ((char *)%0)[(int)%1]=(char)0; '' arr (I# (read# -# 1#)) >>= \ () ->
131 return (StringBuffer a# read# 0# 0#)
133 freeStringBuffer :: StringBuffer -> IO ()
134 freeStringBuffer (StringBuffer a# _ _ _) =
135 _casm_ `` free((char *)%0); '' (A# a#)
137 unsafeWriteBuffer :: StringBuffer -> Int# -> Char# -> StringBuffer
138 unsafeWriteBuffer s@(StringBuffer a _ _ _) i# ch# =
140 _casm_ `` ((char *)%0)[(int)%1]=(char)%2; '' (A# a) (I# i#) (C# ch#) >>= \ () ->
149 currentChar# :: StringBuffer -> Char#
150 currentChar# (StringBuffer fo# _ _ current#) = indexCharOffAddr# fo# current#
152 currentChar :: StringBuffer -> Char
153 currentChar sb = case currentChar# sb of c -> C# c
155 indexSBuffer# :: StringBuffer -> Int# -> Char#
156 indexSBuffer# (StringBuffer fo# _ _ _) i# = indexCharOffAddr# fo# i#
158 indexSBuffer :: StringBuffer -> Int -> Char
159 indexSBuffer sb (I# i#) = case indexSBuffer# sb i# of c -> C# c
161 -- relative lookup, i.e, currentChar = lookAhead 0
162 lookAhead# :: StringBuffer -> Int# -> Char#
163 lookAhead# (StringBuffer fo# _ _ c#) i# = indexCharOffAddr# fo# (c# +# i#)
165 lookAhead :: StringBuffer -> Int -> Char
166 lookAhead sb (I# i#) = case lookAhead# sb i# of c -> C# c
170 moving the start point of the current lexeme.
173 -- moving the end point of the current lexeme.
174 setCurrentPos# :: StringBuffer -> Int# -> StringBuffer
175 setCurrentPos# (StringBuffer fo l# s# c#) i# =
176 StringBuffer fo l# s# (c# +# i#)
178 -- augmenting the current lexeme by one.
179 incLexeme :: StringBuffer -> StringBuffer
180 incLexeme (StringBuffer fo l# s# c#) = StringBuffer fo l# s# (c# +# 1#)
182 decLexeme :: StringBuffer -> StringBuffer
183 decLexeme (StringBuffer fo l# s# c#) = StringBuffer fo l# s# (c# -# 1#)
187 -- move the start and end point of the buffer on by
191 stepOn :: StringBuffer -> StringBuffer
192 stepOn (StringBuffer fo l# s# c#) = StringBuffer fo l# (s# +# 1#) (s# +# 1#) -- assume they're the same.
194 stepOnBy# :: StringBuffer -> Int# -> StringBuffer
195 stepOnBy# (StringBuffer fo# l# s# c#) i# =
197 new_s# -> StringBuffer fo# l# new_s# new_s#
200 stepOnTo# :: StringBuffer -> Int# -> StringBuffer
201 stepOnTo# (StringBuffer fo l _ _) s# = StringBuffer fo l s# s#
203 stepOnUntil :: (Char -> Bool) -> StringBuffer -> StringBuffer
204 stepOnUntil pred (StringBuffer fo l# s# c#) =
208 case indexCharOffAddr# fo c# of
209 ch# | pred (C# ch#) -> StringBuffer fo l# c# c#
210 | ch# `eqChar#` '\NUL'# && c# >=# l# -> StringBuffer fo l# l# l# -- EOB, return immediately.
211 | otherwise -> loop (c# +# 1#)
213 stepOverLexeme :: StringBuffer -> StringBuffer
214 stepOverLexeme (StringBuffer fo l s# c#) = StringBuffer fo l c# c#
216 expandWhile :: (Char -> Bool) -> StringBuffer -> StringBuffer
217 expandWhile pred (StringBuffer fo l# s# c#) =
221 case indexCharOffAddr# fo c# of
222 ch# | pred (C# ch#) -> loop (c# +# 1#)
223 | ch# `eqChar#` '\NUL'# && c# >=# l# -> StringBuffer fo l# l# l# -- EOB, return immediately.
224 | otherwise -> StringBuffer fo l# s# c#
227 scanNumLit :: Int -> StringBuffer -> (Int,StringBuffer)
228 scanNumLit (I# acc#) (StringBuffer fo l# s# c#) =
232 case indexCharOffAddr# fo c# of
233 ch# | isDigit (C# ch#) -> loop (acc# *# 10# +# (ord# ch# -# ord# '0'#)) (c# +# 1#)
234 | ch# `eqChar#` '\NUL'# && c# >=# l# -> (I# acc#, StringBuffer fo l# l# l#) -- EOB, return immediately.
235 | otherwise -> (I# acc#,StringBuffer fo l# s# c#)
238 expandUntilMatch :: StringBuffer -> String -> StringBuffer
239 expandUntilMatch (StringBuffer fo l# s# c#) str =
242 loop c# [] = StringBuffer fo l# s# c#
243 loop c# ((C# x#):xs) =
244 if indexCharOffAddr# fo c# `eqChar#` x# then
251 -- at or beyond end of buffer?
252 bufferExhausted :: StringBuffer -> Bool
253 bufferExhausted (StringBuffer fo l# _ c#) = c# >=# l#
255 emptyLexeme :: StringBuffer -> Bool
256 emptyLexeme (StringBuffer fo l# s# c#) = s# ==# c#
259 prefixMatch :: StringBuffer -> String -> Maybe StringBuffer
260 prefixMatch (StringBuffer fo l# s# c#) str =
263 loop c# [] = Just (StringBuffer fo l# s# c#)
264 loop c# ((C# x#):xs) =
265 if indexCharOffAddr# fo c# `eqChar#` x# then
270 untilEndOfString# :: StringBuffer -> StringBuffer
271 untilEndOfString# (StringBuffer fo l# s# c#) =
275 case indexCharOffAddr# fo c# of
277 case indexCharOffAddr# fo (c# -# 1#) of
279 -- looks like an escaped something or other to me,
280 -- better count the number of "\\"s that are immediately
281 -- preceeding to decide if the " is escaped.
284 case indexCharOffAddr# fo i# of
285 '\\'# -> odd_slashes (not flg) (i# -# 1#)
288 if odd_slashes True (c# -# 2#) then
289 -- odd number, " is ecaped.
291 else -- a real end of string delimiter after all.
292 StringBuffer fo l# s# c#
293 _ -> StringBuffer fo l# s# c#
295 if c# >=# l# then -- hit sentinel, this doesn't look too good..
296 StringBuffer fo l# l# l#
302 untilEndOfChar# :: StringBuffer -> StringBuffer
303 untilEndOfChar# (StringBuffer fo l# s# c#) =
307 case indexCharOffAddr# fo c# of
309 case indexCharOffAddr# fo (c# -# 1#) of
311 case indexCharOffAddr# fo (c# -# 2#) of
312 '\\'# -> -- end of char
313 StringBuffer fo l# s# c#
314 _ -> loop (c# +# 1#) -- false alarm
315 _ -> StringBuffer fo l# s# c#
317 if c# >=# l# then -- hit sentinel, this doesn't look too good..
318 StringBuffer fo l# l# l#
323 untilChar# :: StringBuffer -> Char# -> StringBuffer
324 untilChar# (StringBuffer fo l# s# c#) x# =
328 if indexCharOffAddr# fo c# `eqChar#` x# then
329 StringBuffer fo l# s# c#
334 lexemeToString :: StringBuffer -> String
335 lexemeToString (StringBuffer fo _ start_pos# current#) =
336 if start_pos# ==# current# then
339 unpackCStringBA (copySubStr (A# fo) (I# start_pos#) (I# (current# -# start_pos#)))
341 lexemeToByteArray :: StringBuffer -> _ByteArray Int
342 lexemeToByteArray (StringBuffer fo _ start_pos# current#) =
343 if start_pos# ==# current# then
344 error "lexemeToByteArray"
346 copySubStr (A# fo) (I# start_pos#) (I# (current# -# start_pos#))
348 lexemeToFastString :: StringBuffer -> FastString
349 lexemeToFastString (StringBuffer fo l# start_pos# current#) =
350 if start_pos# ==# current# then
351 mkFastCharString2 (A# fo) (I# 0#)
353 mkFastSubString (A# fo) (I# start_pos#) (I# (current# -# start_pos#))
356 Create a StringBuffer from the current lexeme, and add a sentinel
357 at the end. Know What You're Doing before taking this function
360 lexemeToBuffer :: StringBuffer -> StringBuffer
361 lexemeToBuffer (StringBuffer fo l# start_pos# current#) =
362 if start_pos# ==# current# then
363 StringBuffer fo 0# start_pos# current# -- an error, really.
365 unsafeWriteBuffer (StringBuffer fo (current# -# start_pos#) start_pos# start_pos#)